Pulverized boiler : Radiation to super heaters
 

I am modeling tangentially fired boiler where the super heater are suspended at the top of the combustion zone. Is any one attempted to model radiative heat transfer from flame to these super heaters.

Thanks

Re: Pulverized boiler : Radiation to super heaters
 

Yes.

Re: Pulverized boiler : Radiation to super heaters
 

Thanks Allan for your response

I have model the super-heater (which is suspended over the flame nearly 15-20m apart )as porous volume. I am using P-1 radiation model. By udf I have accounted the local gas radiation to super heater. But Now, how to account the flame radiation? May be view factor method will be useful but I am not sure about the same.

Kindly discuss in brief the methodology adopted for your case.

Re: Pulverized boiler : Radiation to super heaters
 

We have done the superheaters two ways. One, like you, we have done with porous cells using a udf and local temperatures to get the local convective and radiant fluxes. And, two, we have modeled the superheaters as platens using fairly thin solid sections. For each "platen" we would specify the inner temperature, wall thickness, and conductivy, and the radiation properties, and then have Fluent calculate the outer temperature and the P1 radiative flux. This was a more satisfactory approach.

I don't think you will actually have much direct radiation from the flame to the SH 15 to 20 m away. Normally, the gases will be optically thick (with an emissivity of about 0.5). Thus, "radiation" might only propogate for perhaps 3 m before it is reradiated. In theory, the P1 model should be able to track this. If you have a hot pulverized coal flame, the radiation will heat up the gases in the upper furnace, which will in turn radiate more heat to the superheaters.

Some of this you can check by setting up a simple test case with a hot wall radiating to cooler wall with gases of varying optical thickness in-between. How does the distance between walls change the calculated radiation?

Re: Pulverized boiler : Radiation to super heaters
 

Dear Allan,

Thank you for your valuable comments on the problem specified. I will attempt the example given by you and will discuss it separately.

Few queries on second attempt are:

1) As I understand, you had created a bundle of thin slices of solid planes for platen and specified constant wall temp condition, emissivity and wall thickness to the same. ( This will look like actual suspended platen SH in model)

-But did you ensured that the heat transfer surface area remains same as actual. ( I am getting 21% less area by this method)

-This will also affect the mesh quality in this zone or increase the number of cells to large extent.

-How much difference was observed in heat flux when both options were compared.

-Did you specify the constant wall temp based on average of i/p and o/p steam temperature.

-Have u published/presented paper on the same, can you pl send me soft copy of the same.

2) One more problem I was facing:

I am simulating 200 MWe tangentially coal fired boiler.

the models are :RNG -k-e, P-1, Species transport with volumetric reactions , diffusion controlled, constant rate devolatilization.

The problem is temp shoots up to 3000K. I have checked heat balance it is 6% off. The actual operating temp is < 1800K.

Tried DO as but results are same.

Pl. advice

Re: Pulverized boiler : Radiation to super heaters
 

In answer to questions: - for surface area: are you taking the circumferential area of each superheater tube? In boiler design calculations, the actual, projected, or effective areas can be different and are accounted for by fouling efficiency factors. In a PC boiler, depending on coal, fouling factors can be up to 50%. But, wasn't your question about heat fluxes to the SH from the PC burner flame? Does the radiation actually travel that far. - Yes, the mesh will have to account for SH platen shapes and likely you will not have enough grid resolution to model convective heat transfer with any accuracy, but again, your question was on radiation. - We haven't used the porous cell and platen approaches on the same boiler so difficult to compare. Anyway, we always have had measured flue gas temperatures downstream of the superheater, and the superheater heat transfer is tuned to match temperature. For design of replacement superheaters, we don't use CFD as the primary tool. - Wall temperatures are some weighted average. - There are papers out there on our superheater designs (google Jansen Combustion) but again, CFD is not the primary tool.

On the temperature: - I'm guessing you have already made the gas heat capacity a function of temperature and composition. This will change the temperature quite a bit. - Are you using the WSGGM for gas absorptivity/emissivity? Do you add a factor for luminosity of the PC particles? This would decrease temps. a little. What about your reaction rates? Maybe too sensitive to temperature? And volatilization? You could limit the temperatures to less than 3000K and see if it steadies out. (Radiation is slow so cells will heat up and reactions will be fast so they may not cool back down.) - In our cases, the energy balance eventually works out. If you are using ver6.3, you can (easily) look at the sensible heat balance as well. The outlet temperature should take a long time to converge.

Re: Pulverized boiler : Radiation to super heaters
 

Dear Allan,

>> I'm guessing you have already made the gas heat capacity a function of temperature and composition. This will change the temperature quite a bit.

-YES

>> Are you using the WSGGM for gas absorptivity/emissivity?

-YES

>>Do you add a factor for luminosity of the PC particles?This would decrease temps. a little.

-As I understand it is particle emissivity, = 0.9

>>What about your reaction rates? Maybe too sensitive to temperature? And volatilization?

-Devolatilization: Constant rate (12 s-1) Char oxidation: Diffusion controlled C+ O2 ==CO2 Volatile: Finite rate/EDC (very fast kinetics)

The above models are assumed in absence of data on subbituminous medium volatile coal. I guess the constant rate Devolatilization may be leading to instatenous gas phase reactions which may be leading to high localize temperature. Can you comment on this?

>> You could limit the temperatures to less than 3000K and see if it steadies out. (Radiation is slow so cells will heat up and reactions will be fast so they may not cool back down.)In our cases, the energy balance eventually works out. If you are using ver6.3, you can (easily) look at the sensible heat balance as well. The outlet temperature should take a long time to converge.

-I will do this part and communicate back.

Re: Pulverized boiler : Radiation to super heaters
 

So the gas absorption coefficient is set with the wsggm. You can look at this to back out the effecitive emissivity and compare with literature values. But what about the scattering coefficient? This is a different factor in the P1 model than the particle emissivity. As a simple model, you could increase the scatting coefficient as a function of temperature to make the flame more luminous. There must be literature data, say from Hottel and Sarofim, that would let you tune in this parameter.

If you want some literature values say for carbon monoxide reation rates, look for papers from Reaction Engineering International.

Re: Pulverized boiler : Radiation to super heaters
 

Dear Allan,

This is about the radiative heat transfer to platen.

>> Some of this you can check by setting up a simple test case with a hot wall radiating to cooler wall with gases of varying optical thickness in-between. How does the distance between walls change the calculated radiation?

-I performed two experiments with a closed box similar to furnace and a platen suspended from the top which was defined as wall. It was 2-D configuration with right side, left side and bottom wall. The bottom wall was specified as 2000K and the other walls were adiabatic. The platen was specified as 600K. The fluid domain was initialized with CO2-16%, H20-10%, O2 3% and rest N2 (mass frac). The radiation model was P-1. In case 1, the distance between platen and bottom wall was 10m and in case 2 as 20m.

The absorption coeff. calculated by WSGGM was ranging from 0.4-2.5 m-1. The scattering coefficient was 0.6 m-1. It was observed that in both cases the heat transferred to platen was > 95%. I tried to estimated the change in intensity, I = Io * exp (-a*L) for the a=0.5m-1. The I/Io comes to be 0.000045. i.e. very negligible absorption of radiation intensity.

Does this means that the radiation from flame can travel through a thick medium of flue gas and heat the platen. Or the WSGGM is estimating very high values of absorption coefficient.